iterative phase noise compensation

该研究针对多用户MIMO系统中由相位噪声引起的问题，提出了一种基于变分贝叶斯框架的方法进行联合相位噪声估计和数据检测，有效提升了通信系统的性能。

Conventional expectation propagation (EP) detection suffers severe performance degradation in strongly intersymbol-interference (ISI) channels due to inaccurate initial linear minimum mean-square-error (LMMSE) estimates. Method: This paper proposes a transform-domain iterative message-passing detection framework. It incorporates channel-shortening filtering to compress the time-domain impulse response and jointly optimizes linear EP and low-complexity BCJR detection in the transform domain. A deliberate initialization mismatch strategy is introduced to accelerate convergence, and symbol-level nonlinear mapping is replaced by a BCJR detector with controllable state complexity to enhance robustness. Contribution/Results: The method significantly improves message covariance estimation and prior matching. Experiments on Proakis-C and measured wireless channels demonstrate up to 6 dB bit-error-rate gain over baseline schemes, support 2 bits per channel use, and achieve superior performance–complexity trade-offs.

This work proposes a novel Bayesian-optimal iterative signal recovery algorithm for multiuser linear Gaussian communication systems with randomly right unitarily invariant precoding. Built upon the Orthogonal Approximate Message Passing (OAMP/VAMP) framework, the method achieves efficient iterative updates through interpolation between Expectation Propagation (EP) and OAMP, enabling, for the first time, Bayesian-optimal reconstruction of signals with non-separable priors. The authors innovatively introduce a disorder-averaging technique combined with the replica-symmetric ansatz to establish a finite-sample high-dimensional analysis of the algorithm. Theoretical analysis demonstrates that the proposed algorithm attains Bayesian optimality in the large-system limit and aligns precisely with replica-symmetric predictions, exhibiting superior performance in multiuser communication scenarios.

This paper addresses blind joint channel estimation and symbol detection for time-varying linear intersymbol interference (ISI) channels without pilot symbols. Method: We propose an EM-BP joint iterative framework that integrates factor graph modeling, expectation-maximization (EM), and belief propagation (BP). Key innovations include a data-driven momentum-enhanced BP update rule and a learnable EM parameter scheduling strategy, both optimized offline via small-sample training. Contribution/Results: Compared to conventional coherent BP detection, the proposed method achieves superior bit error rate (BER) performance at high signal-to-noise ratios (SNRs) while significantly reducing computational complexity. Numerical experiments demonstrate robust blind detection capability and an excellent trade-off between performance and complexity.

Existing channel estimation methods fail for extra-large-scale MIMO (XL-MIMO) under the joint effects of near-field (NF) propagation, double-bandwidth (DB) operation, and spatial non-stationarity (SnS), where conventional channel sparsity structures collapse. Method: This paper proposes a two-stage decoupling framework integrated with a three-layer Bayesian inference mechanism. It introduces a novel structured sparse prior comprising angular-domain block sparsity, spatially non-stationary modeling, and subchannel signal decoupling—specifically tailored to spherical-wave NF channels—and designs the first three-layer generalized approximate message passing (TL-GAMP) algorithm for such channels. Results: The proposed method achieves stable convergence across diverse channel regimes—including NF-SnS, NF-stationary (SS), and far-field (FF)-SS—and reduces estimation error by over 30% while maintaining near-linear computational complexity. It significantly enhances both accuracy and scalability of broadband XL-MIMO channel estimation.

This work addresses the impractical computational complexity of optimal detectors, such as maximum a posteriori (MAP), in large-scale MIMO systems under PSK modulation, where complexity grows exponentially with the modulation order. To overcome this limitation, the authors propose a novel belief propagation detector grounded in directional statistics, which introduces the von Mises distribution into the message-passing framework for the first time. By continuously relaxing PSK symbols onto the unit circle and parameterizing messages accordingly, the method achieves a sparse representation whose complexity is independent of the modulation order. The proposed detector significantly reduces computational overhead, accommodates imperfect channel state information, and demonstrates superior performance over Gaussian approximation-based detection algorithms across various PSK modulations and channel conditions.

This study addresses the challenge of multi-target MIMO sensing under unknown angles and interfering reflection coefficients by proposing a progressive Bayesian sensing framework. The framework leverages variational Bayesian inference to efficiently approximate high-dimensional posterior distributions, replacing exponential numerical computations with polynomial complexity. Furthermore, it iteratively updates priors using posteriors to optimize transmit beamforming, thereby minimizing the Posterior Cramér-Rao Bound (PCRB). This research significantly reduces system computational complexity while validating the effectiveness of the proposed framework in progressively refining multi-target sensing performance.

This work addresses the high computational complexity of message-passing detectors in massive MIMO systems by proposing a low-complexity receiver framework based on orbital priors. By relaxing discrete symbol priors into mixed discrete–continuous densities and leveraging an orbital prior decomposition, the symbol distribution is compressed into 3L real-valued scalars, yielding three closed-form denoisers—OBD, OGD, and OPD—that reduce per-iteration complexity to O(1). Combining the Jacobi–Anger expansion, state evolution analysis, and optimal transport theory, the proposed method asymptotically achieves capacity (log₂M), eliminates error floors, attains an MMSE dimension of d = 1/2, and establishes a Wasserstein-distance bound linking constellation ring geometry to achievable rates.

该论文研究了在MIMO系统中通过使符号遵循Maxwell-Boltzmann分布简化概率整形解调问题，提出的方法等效于对均匀星座进行简单的预处理步骤。

This work addresses the challenge that approximate message passing (AMP) with random initialization struggles to effectively recover signals within a fixed time in noiseless phase retrieval. By characterizing the algorithm’s dynamical trajectory through Gaussian decomposition and combining refined long-time error control with state evolution analysis under the generalized AMP framework, the authors rigorously establish—for the first time—that weak recovery is achievable when the sampling rate δ exceeds 1/2, and arbitrarily precise recovery is attained in O(log n) iterations when δ > 1.13, surpassing the limitations of conventional state evolution theory. Moreover, for δ ∈ (0.5, 1.13), the algorithm reliably converges to a finite fixed point, thereby confirming the efficacy of random initialization across this regime.